Summary Dental caries is the most significant problem in oral health area. Early detection of caries is highly desired because decays can be reversed by remineralization. When detected later, caries result in costly treatments such as fillings, bridge, crown or dental implants, amounting to thousands of dollars per tooth. Therefore, the reliable diagnosis of decay at the early onset would have a significant impact by curbing the progression of the disease before the advanced stage of tooth decay and, in turn, alleviate the associated dental-care costs. Because early onset lesions are elusive to current technologies, there is great emphasis on developing new technologies for early diagnosis of caries. Raman imaging is promising but it needs to be lower cost and customized for the dental operatory. Raman can detect caries by visualizing changes in the mineralization of decaying enamel. The speed of existing Raman modalities (point by point scan etc.) does not allow a systematic survey of the oral space, and the Raman microscopes are too bulky and complex to be used in the dental operatory. A fiber optic Raman- imaging tool is essential to enable clinical utilization of Raman. The major aim of the proposed pilot study is to develop a Raman imaging device for early detection of caries. Aim 1 will integrate a hand- held fiber optic Raman imager for dental diagnostics. This imager is a portable and automated platform that is an innovative combination of microarray based wide-field excitation and 2D- signal collection optics to generate mineralization distributions of the occlusal, frontal and posterior aspects of the teeth. The system will generate full-field images of teeth surface finding lesions as least 100 m size in the oral space within 5 minutes. Aim 2 will validate and refine the performance of the imager towards the detection of caries. The limits of detection (size of the lesion and the degree of mineralization) of the system will be assessed and improved by artificially induced lesions of controlled size and severity. The imager will then be tested blindly on sound teeth samples and those with naturally occurring caries lesions to determine its sensitivity and specificity against the gold standard undemineralized histology.